In an MSVC build, not exporting BUILD_OPT to the Mozilla build
machinery causes the produced DLLs to use the debug CRT. The exact
mechanism is a bit of a mystery, and I didn't feel like spending too
much time trying to understand it.

Using the debug CRT is confusing and wrong. Nothing in LO otherwise
uses it. It also makes testing a build much harder for me at least, as
I do that in a fairly pristine virtual machine with no MSVC debugging
runtime available. (The normal CRT is bundled in the LO installer.)

Change-Id: I27f774d92a3986d40162c870202bcdddd94aa7c6

Change-Id: Ifd59198a2d09e26377267c8b509c5340d83240e3

Change-Id: I6223cd12ab539cd19555c0b23f2bce73f519a31f

I always wanted to do this.  The new name reflects what it really does
better than the old one.

Change-Id: I3c90cce06d482f1453936ff3916eef9663bb417b

Change-Id: Idd3ca250ea51c9e17b4566febf13d9c8ca9c6a8f

There is no point of having hasCache() method which always returns true.

Change-Id: I309d3c33899cdab2049af5199ae8d41f7cbd5f99

They were correct up until IBM re-worked a whole bunch of stuff during
the OOo 3.2 (?) development cycle.

Change-Id: Ia9a9e30287f6c8451a6cfa98f3515d703e3c67cc

Change-Id: I10f8d534d664c5f27a32214c00aeee890d993d18

Perhaps this will make valgrind happy?

Change-Id: I11f965d13f8d71b2ab9f9624367fd70713644f10

Change-Id: Iae3eceb38a6f9bc53805445799534ec8ee44a9a9

Change-Id: I20bbd2d5ed7ba3c84ee33544f4a1a39c02729c23

When removing an entry from the tree list, we need to

1) first remove the entry from the list,
2) broadcast its removal with the entry instance, then
3) finally delete the entry itself at the very end.

in this exact order, or else interesting stuff would ensue.

Change-Id: If42b141921ffe4ed36dce31f93a4a084204815bf

Change-Id: I48cd5988056c8dee10241d5ba3b469ae7e5191b9

Change-Id: I9bb589e8a349a1404f9c09a60a82e824918744fa

"make -jyes" considered useless

Change-Id: I044e95f76e8ea961e781e75c4b6923e700e0d375

Change-Id: I815ef8abaf4cd998e7b91fbadad56ddf0a7087ba

Change-Id: Ieba85dd0093c9f704090075d2eb75f1eefdd72de

...introduced with 75bc673b "Travelled MIME type
through UCB for CMIS UCP to use it when creating documents" and causing crashes
in sc_unoapi test.  Many SfxMedium instances appear to have a null filter, so
this workaround at least fixes the crashes.  No idea what a good fix should look
like, so will ask Cedric to review his commit.

Change-Id: I27180265b6c79c7ed5f459b03c9e188ff6a0f56d

Change-Id: I91ded7801f4959f2dac7b5b3e4c8d1f3929f4ed9

Change-Id: I8474e5da02036f065057a5610def6701db4cedc7

Change-Id: I431c0615de56ddc0fa2c0ea06567d32419d29440

Change-Id: I655e8e15c7f7a8c292b3a1820ee48c29e847d05a

Change-Id: I98aa2bcb90385096915c9cddea6a9e68ab95d29c

Change-Id: Id54d9bd2963573c4b8e77de0dfafe33ea5882d32
Signed-off-by: Marcos Paulo de Souza <marcos.souza.org@gmail.com>
Reviewed-on: https://gerrit.libreoffice.org/947Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

use plain copy instead of git-new-workdir when using --with-linked-git
configure switch (point it to your master copy of the core-repo)
also remove obsolete bin/create_bootstrap_links, bin/mkworkdir and
bin/repo-list[.in]

Change-Id: I37028badf2c46a24a3d628e80476efe4dd365d06
Reviewed-on: https://gerrit.libreoffice.org/912Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

Change-Id: Ibb9bbf018e4e9c1487ccdf6505a2a15d7c24eb24

Project: help  eab79dba5c42eaeb6ecac4a9b6eeb0e6500f3301
update macro dialog help ids for .ui

Change-Id: Ia283f2e70fa623d0155b6de51ad5441163505cc0
Signed-off-by: Marcos Paulo de Souza <marcos.souza.org@gmail.com>
Reviewed-on: https://gerrit.libreoffice.org/945Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

MSVC warns about the anonymous struct, and the way we write and read
from the union is undefined.

Change-Id: I683f54d90393a744e5c94857cd87800bcc8f9ffe

These cause valgrind warnings in sc_unoapi.
(regression from d2b49b5b)

Change-Id: If3128410bfabc23ff85c24156a8602379b8b7edd

Change-Id: Id488ffa6a6b01937fe32e4c80af45c40ebf673ec

Change-Id: Id7f1471b49af52e6f6b0515ccd1fe8e12c50d9b5

Change-Id: I4cb64ae2b2f2dbf643e38c5208eb759f265acafd

LO builds GLib, gdk-pixbuf and Pango using their ancient nmake-based
build mechanisms (that use the make.msc from GLib's build/win32
directory). This compiles against the debugging runtime (the -MDd
switch, msvcr90d.dll) if there is a DEBUG environment variable. We
don't want that. None of LO's own DLLs should use the debugging
runtime even in an --enable-debug build, and the debugging runtime
doesn't get included in the installer.

Change-Id: I04d2dcc0d5bb245af4afac976b6c1fecbf2095e1

Due to an insane amount of ridiculous "user-defined" number formats the
internal table may overflow during import, stave off and increase the
number of available slots to another arbitrary limit.

...in favor of existing new-style configuration::theDefaultProvider singleton.

Theoretically, ConfigurationProvider instances can be created with specific
Locale and EnableAsync arguments, but this is hardly used in practice, and thus
effectively all uses of the ConfigurationProvider service use the
theDefaultProvider instance, anyway.

theDefaultProvider is restricted to the XMultiServiceFactory interface, while
ConfigurationProvider also makes available XComponent.  However, dispose must
not be called manually on theDefaultProvider singleton anyway, and calls to
add-/removeEventListener are so few (and in dubious code that should better be
cleaned up) that requiring an explicit queryInterface does not really hurt
there.

This commit originated as a patch by Noel Grandin to "Adapt
configuration::ConfigurationProvider UNO service to new style [by creating] a
merged XConfigurationProvider interface for this service to implement."  It was
then modified by Stephan Bergmann by deprecating ConfigurationProvider instead
of adding XConfigurationProvider and by replacing calls to
ConfigurationProvider::create with calls to theDefaultProvider::get.

Change-Id: I9c16700afe0faff1ef6f20338a66bd7a9af990bd

After 047bc546 it is no longer triggered
by rsvg alone, so I guess this is the right way to get system glib on migw again.

Change-Id: Ie8a6c2a8850ad3dd6cd2be878aedfbf81d2473f7

Change-Id: I7c7f951b352ed2e9ef7604a5be324223f21e96dd

Change-Id: I4e83afa3d164ba0ae8640f4b5f2712b88c1e305f